To evaluate the safety and efficacy of E5 (Xomen-E5), a murine monoclonal immunoglobulin M antibody, in reducing mortality in patients with serious Gram-negative infections. Phase II, single-site study. Randomized, double-blind, placebo-controlled study. Surgical, neurosurgical and medical ICUs, comprising approximately 30 beds in a multidisciplinary university hospital. Patients with clinical evidence of serious infection admitted to the ICU for monitoring of vital signs and for intensive care nursing. Of the 39 patients enrolled in the study, 23 had documented Gram-negative infection. Patients suspected of having life-threatening Gram-negative infections received one of two doses of E5 or placebo. Safety and efficacy were assessed by survival on days 3, 7, and 21, appearance of adverse reactions, development of antimurine antibodies, and effects on BP, urine output, WBC count, and temperature. Mortality rate from Gram-negative infection 3 days after the last drug (or placebo) infusion was two (22%) of nine deaths in the placebo group compared with 0 of nine for E5 2.5 mg/kg and 0 of five for E5 7.5 mg/kg. At 21 days after therapy, three patients treated with E5 had died. Only one of these three deaths resulted from infection. Eight of 15 E5 patients tested had immunoglobulin G antimurine antibodies by 3 wks after therapy began, but all were asymptomatic. E5 was well tolerated and may have the potential to reduce early morbidity and the mortality rate in seriously ill patients with Gram-negative infections. Results from larger phase III studies are needed to confirm these findings.